Bonus 1 - A library: the recorded walkthrough

Recorded with Claude Code 2.1.289, from the lab's instructions. Results are shortened; your agent's answers will differ in wording.

Step 1 - Start the project next to the clone

The participant runs uv init --lib demoshop-library:

Initialized project `demoshop-library` at `<repo>`

The participant runs uv add "robotframework==7.5" "robotframework-assertion-engine==5.0.1" "robotframework-pythonlibcore==4.6.0" "requests==2.34.2" uv add --dev pytest:

 ~ demoshop-library==0.1.0 (from file://<repo>)
 + iniconfig==2.3.0
 + packaging==26.3
 + pluggy==1.6.0
 + pygments==2.21.0
 + pytest==9.1.1

Step 2 - Save the references

The participant runs `mkdir references curl -s http://localhost:9090/openapi.json -o references/demoshop-openapi.json curl -sL https://raw.githubusercontent.com/MarketSquare/AssertionEngine/main/README.md -o references/assertionengine-readme.md

Keywords to imitate, and AssertionEngine's code, from the workshop's environment

cd ../ai-engineering-robotframework uv run robotcode libdoc Browser show "Get Text" > ../demoshop-library/references/browser-get-text.md uv run robotcode libdoc Browser show "Get Element Count" > ../demoshop-library/references/browser-get-element-count.md uv run --no-sync python -c "import shutil, assertionengine.assertion_engine as m; shutil.copy(m.file, '../demoshop-library/references/assertion_engine.py')" cd ../demoshop-library`:

(no output)

The participant runs ls -l references | cut -c30-:

Oct  6 19:57 assertion_engine.py
Oct  6 19:57 assertionengine-readme.md
Oct  6 19:57 browser-get-element-count.md
Oct  6 19:57 browser-get-text.md
Oct  6 19:57 demoshop-openapi.json

Step 3 - Write AGENTS.md

In the lab you write this file yourself. For the reference, the rehearsal lets the agent draft it from the lab's own list, and the transcript says so.

Prompt:

Write AGENTS.md for this project from this list, as the sections Toolstack, References, Concepts, Examples and Specification, short, and nothing else:

Toolstack:

Python 3.12 with uv;

dependencies at run time: Robot Framework 7.5, AssertionEngine, PythonLibCore and requests, pinned in pyproject.toml;

unit tests with uv run pytest;

Robot tests in atest/, against the local shop at http://localhost:9090, run with uv run robot --outputdir results atest;

the keyword documentation is generated with uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html;

packaging with uv build.

References:

references/demoshop-openapi.json, the API of the shop version under test;

the User Guide's Creating test libraries and library scope, for Robot Framework 7.5.

Concepts:

every Get keyword takes assertion_operator, assertion_expected and message, and checks with AssertionEngine's verify_assertion (references/assertion_engine.py, references/assertionengine-readme.md);

the library is built on PythonLibCore;

it is imported with url and an optional space, which it sends as X-Workshop-Space;

each library instance keeps one cart.

Examples: references/browser-get-text.md and references/browser-get-element-count.md: keyword names, arguments and documentation to imitate.

Specification: OpenSpec, under openspec/changes/.

The agent runs ls -a && ls references src && cat pyproject.toml

.
..
.git
.gitignore
.python-version
.venv
README.md
pyproject.toml
references
src
uv.lock
references:
... (32 more lines)

The agent writes <repo>/AGENTS.md

File created successfully at: <repo>/AGENTS.md (file state is current in yo

Agent:

I created AGENTS.md in the project root with the five sections you listed: Toolstack, References, Concepts, Examples and Specification. It contains only the content you gave me.

I checked the paths it points to. The files under references/ exist, and the four run-time dependencies are pinned in pyproject.toml. Three locations it names don't exist yet: atest/, docs/ and openspec/.

The participant runs cat AGENTS.md:

# AGENTS.md

## Toolstack

- Python 3.12 with uv.
- Run-time dependencies: Robot Framework 7.5, AssertionEngine, PythonLibCore and requests, pinned in `pyproject.toml`.
- Unit tests: `uv run pytest`.
- Robot tests live in `atest/` and run against the local shop at `http://localhost:9090`:
  `uv run robot --outputdir results atest`.
- Keyword documentation:
  `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html`.
- Packaging: `uv build`.
... (25 more lines)

Step 4 - Set up OpenSpec

The participant runs openspec init --tools claude:

Learn more: https://github.com/Fission-AI/OpenSpec
Feedback:   https://github.com/Fission-AI/OpenSpec/issues

- Setting up Claude Code...
[32m✔[39m Setup complete for Claude Code

In the lab you write this file yourself. For the reference, the rehearsal lets the agent draft it from the lab's own list, and the transcript says so.

Prompt:

In openspec/config.yaml, uncomment context: and write there, in a few lines each, the five points of AGENTS.md: toolstack, references, concepts, examples and specification. Change nothing else in the file.

The agent reads <repo>/openspec/config.yaml

The agent edits <repo>/openspec/config.yaml

The file <repo>/openspec/config.yaml has been updated successfully. (file s

The agent runs (python3 -c "import yaml,sys; d=yaml.safe_load(open('openspec/config.yaml')); print(sorted(d)); print(d['context'])" 2>&1 || uv run --with pyyaml python -c "import yaml; d=yaml.safe_load(open('openspec/config.yaml')); print(sorted(d)); print(d['context'])")

pyenv: version `3.12' is not installed (set by <repo>/.python-version)
['context', 'schema']
Toolstack:
- Python 3.12 with uv. Run-time dependencies: Robot Framework 7.5, AssertionEngine,
  PythonLibCore and requests, pinned in pyproject.toml.
- Unit tests: `uv run pytest`. Robot tests live in atest/ and run against the local shop
  at http://localhost:9090: `uv run robot --outputdir results atest`.
- Keyword documentation: `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html`.
  Packaging: `uv build`.

References:
- references/demoshop-openapi.json: the API of the shop version under test.
... (18 more lines)

Agent:

I uncommented context: in openspec/config.yaml and replaced the three sample lines under it with the five points from AGENTS.md: Toolstack, References, Concepts, Examples and Specification. Each is a short labelled section. I checked the file with a YAML parser: it parses and has two keys, schema and context.

Nothing else in the file changed. One thing reads a little oddly: the # Example: comment above context: is still there. It belonged to the sample I replaced. I left it because you asked me not to change anything else, but you may want to delete that one line.

The participant runs sed -n '1,30p' openspec/config.yaml:

schema: spec-driven

# Project context (optional)
# This is shown to AI when creating artifacts.
# Add your tech stack, conventions, style guides, domain knowledge, etc.
# Example:
context: |
  Toolstack:
  - Python 3.12 with uv. Run-time dependencies: Robot Framework 7.5, AssertionEngine,
    PythonLibCore and requests, pinned in pyproject.toml.
  - Unit tests: `uv run pytest`. Robot tests live in atest/ and run against the local shop
    at http://localhost:9090: `uv run robot --outputdir results atest`.
... (18 more lines)

Prompt:

Which instruction files did you load at the start of this session? Name each file with its path.

Agent:

Only one instruction file was loaded at the start of this session:

<repo>/AGENTS.md is the project instruction file, checked into the codebase.

No CLAUDE.md was loaded, either at the project level or the user level. No memory index (MEMORY.md) was loaded either.

AGENTS.md mentions other files under references/ (for example demoshop-openapi.json, assertion_engine.py and browser-get-text.md). Those were not loaded. They're only listed there as files to consult when needed.

Step 5 - Propose the first slice

Prompt:

/opsx:propose A Robot Framework library, demoshop_library.DemoShopLibrary, for the DemoShop's REST API, built on PythonLibCore, and imported with url and an optional space. A first slice of keywords: Get Product Count, Get Product Price (a product id), Add Product To Cart (a product id and an optional quantity), Get Cart Item Count and Get Cart Total. Every Get keyword takes assertion_operator, assertion_expected and message, and checks with AssertionEngine, so that Get Product Price 1 == 249.99 works as Browser's Get Text does. Each library instance keeps one cart. Unit tests with pytest replace the HTTP layer; Robot tests in atest/ run every keyword against the local shop.

The agent runs openspec list --json; echo "---"; openspec context --json

{
  "changes": [],
  "root": {
    "path": "<repo>",
    "source": "nearest"
  }
}
---
{
  "root": {
    "path": "<repo>",
    "source": "nearest",
... (5 more lines)

The agent runs ls -la openspec openspec/* ; cat openspec/config.yaml 2>/dev/null || cat openspec/config.yml 2>/dev/null; echo "---"; find . -path ./.venv -prune -o -path ./.git -prune -o -type f -print | head -100

-rw-r--r-- 1 user user 2282 Oct  6 19:58 openspec/config.yaml

openspec:
total 20
drwxr-xr-x 4 user user 4096 Oct  6 19:58 .
drwxr-xr-x 8 user user 4096 Oct  6 19:58 ..
drwxr-xr-x 3 user user 4096 Oct  6 19:58 changes
-rw-r--r-- 1 user user 2282 Oct  6 19:58 config.yaml
drwxr-xr-x 2 user user 4096 Oct  6 19:58 specs

openspec/changes:
total 12
... (95 more lines)

The agent runs cat src/demoshop_library/__init__.py; echo "--- pyproject"; cat pyproject.toml; echo "--- README"; cat README.md; echo "--- gitignore"; cat .gitignore; cat .python-version

def hello() -> str:
    return "Hello from demoshop-library!"
--- pyproject
[project]
name = "demoshop-library"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
authors = [
    { name = "Workshop rehearsal" }
]
requires-python = ">=3.12"
... (28 more lines)

The agent reads <repo>/references/browser-get-text.md

The agent reads <repo>/references/browser-get-element-count.md

The agent reads <repo>/references/assertion_engine.py

The agent reads <repo>/references/assertionengine-readme.md

The agent runs wc -c references/demoshop-openapi.json && python3 -c " import json d=json.load(open('references/demoshop-openapi.json')) print(json.dumps({k:v for k,v in d.items() if k not in ('paths','components')}, indent=1)) for p,ops in d['paths'].items(): for m,o in ops.items(): if isinstance(o, dict): print(m.upper(), p, '-', o.get('summary') or o.get('operationId')) "

Exit code 1
8497 references/demoshop-openapi.json
pyenv: version `3.12' is not installed (set by <repo>/.python-version)

Agent:

It's small enough to read directly.

The agent reads <repo>/references/demoshop-openapi.json

The agent runs cd /tmp && for p in /health /api/products/ /api/products/1 /api/products/999999; do echo "=== GET $p"; curl -s -m 5 -i "http://localhost:9090$p" | head -c 3000; echo; done

=== GET /health
HTTP/1.1 200 OK
date: Tue, 06 Oct 2026 18:00:15 GMT
server: uvicorn
content-length: 33
content-type: application/json
cache-control: private, no-store

{"status":"ok","version":"0.3.0"}
=== GET /api/products/
HTTP/1.1 200 OK
date: Tue, 06 Oct 2026 18:00:15 GMT
... (25 more lines)

The agent runs cd /tmp && echo "=== products tail"; curl -s -m 5 http://localhost:9090/api/products/ | tail -c 700; echo; echo "=== products keys/count (jq?)"; which jq; curl -s -m 5 http://localhost:9090/api/products/ | grep -o '"id":[0-9]*' | wc -l; echo "=== GET /api/cart/ no session"; curl -s -m 5 -i http://localhost:9090/api/cart/; echo; echo "=== GET /api/cart/ fresh session"; curl -s -m 5 -i -H "X-Session-ID: opsx-probe-readonly-1" http://localhost:9090/api/cart/; echo; echo "=== products with space header"; curl -s -m 5 -i -H "X-Workshop-Space: opsx-probe" http://localhost:9090/api/products/ | head -c 600; echo

=== products tail
"Health","inventory":60,"rating":4.7,"review_count":321},{"id":9,"name":"Summit Trail Shoes","sku":"SUM-SHO-009","price":149.0,"description":"Trail shoes with adaptive grip pattern, live weather alert
=== products keys/count (jq?)
/usr/bin/jq
12
=== GET /api/cart/ no session
HTTP/1.1 200 OK
date: Tue, 06 Oct 2026 18:00:20 GMT
server: uvicorn
content-length: 48
content-type: application/json
cache-control: private, no-store
... (21 more lines)

The agent runs cd /tmp && S="opsx-probe-$(date +%s%N)"; H="X-Session-ID: $S"; U=http://localhost:9090 echo "=== inventory before"; curl -s $U/api/products/1 | jq -c '.product.inventory' echo "=== add 1 x2"; curl -s -i -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":1,"quantity":2}' $U/api/cart/items; echo echo "=== add 2 x1 (default qty)"; curl -s -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":2}' $U/api/cart/items; echo echo "=== add 1 x1 again (merge?)"; curl -s -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":1,"quantity":1}' $U/api/cart/items; echo echo "=== cart"; curl -s -m 5 -H "$H" $U/api/cart/ | jq . echo "=== qty 21"; curl -s -i -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":1,"quantity":21}' $U/api/cart/items | sed -n '1p;$p'; echo echo "=== unknown product"; curl -s -i -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":999999}' $U/api/cart/items | sed -n '1p;$p'; echo echo "=== qty 0"; curl -s -i -m 5 -H "$H" -H 'Content-Type: application/json' -d '{"product_id":1,"quantity":0}' $U/api/cart/items | sed -n '1p;$p'; echo echo "=== clear"; curl -s -i -m 5 -X DELETE -H "$H" $U/api/cart/ | sed -n '1p;$p'; echo echo "=== cart after clear"; curl -s -m 5 -H "$H" $U/api/cart/; echo echo "=== inventory after"; curl -s $U/api/products/1 | jq -c '.product.inventory'

=== inventory before
25
=== add 1 x2
HTTP/1.1 200 OK
date: Tue, 06 Oct 2026 18:00:38 GMT
server: uvicorn
content-length: 174
content-type: application/json
cache-control: private, no-store

{"session":"opsx-probe-1791309638941688706","items":[{"product_id":1,"name":"Aurora Neural Headphones","quantity":2,"unit_price":249.99,"total_price":499.98}],"total":499.98}
=== add 2 x1 (default qty)
... (41 more lines)

Agent:

Probe done. The cart is cleared and inventory is unchanged at 25. Next I'll check how Robot Framework 7.5 and PythonLibCore behave, using whatever copies are installed locally.

The agent runs ls -a; which uv; uv --version 2>/dev/null; ls ~/.cache/uv 2>/dev/null | head; find / -path /proc -prune -o $ -name typeconverters.py -path "*robot/running/arguments*" $ -print 2>/dev/null | head; find / -path /proc -prune -o $ -type d -name robotlibcore -o -type d -name assertionengine $ -print 2>/dev/null | head

Command did not complete within its 120s timeout and was moved to the background (ID: bot6irkjt). Output is being written to: <scratch>

The agent reads <scratch>

The agent uses ToolSearch with {"query": "select:TaskStop", "max_results": 1}

The agent uses TaskStop with {"task_id": "bot6irkjt"}

{"message":"Successfully stopped task: bot6irkjt (ls -a; which uv; uv --version 2>/dev/null; ls ~/.cache/uv 2>/dev/null | head; find / -path /proc -prune -o \\( -name typeconverters.py -path \"*robot/

The agent runs ls .venv/lib/*/site-packages/ 2>/dev/null; ls -la .venv/bin | head -30

_pytest
_virtualenv.pth
_virtualenv.py
assertionengine
certifi
certifi-2026.7.22.dist-info
charset_normalizer
charset_normalizer-3.5.2.dist-info
demoshop_library-0.1.0.dist-info
demoshop_library.pth
idna
idna-3.20.dist-info
... (43 more lines)

Agent:

The project virtualenv already has everything installed. I'll run a quick experiment outside the project to see how Robot converts assertion_expected under different type hints.

*The agent runs `cd && mkdir -p rf-probe && cd rf-probe && cat > Probe.py <<'EOF' from typing import Any from assertionengine import AssertionOperator, verify_assertion, float_str_verify_assertion, int_str_verify_assertion from robotlibcore import DynamicCore, keyword

class Probe(DynamicCore): ROBOT_LIBRARY_SCOPE = "TEST" def init(self, url: str, space: str | None = None): self.url, self.space = url, space DynamicCore.init(self, [])

@keyword
def union_float_str(self, assertion_expected: float | str | None = None):
    return f"{type(assertion_expected).__name__}:{assertion_expected!r}"

@keyword
def any_expected(self, assertion_expected: Any = None):
    return f"{type(assertion_expected).__name__}:{assertion_expected!r}"

@keyword(tags=("Assertion", "Getter"))
def get_price(self, assertion_operator: AssertionOperator | None = None, assertion_expected: Any = None, message: str | None = None) -> float:
    return float_str_verify_assertion(249.99, assertion_operator, assertion_expected, "Price", message)

@keyword
def get_price_plain(self, assertion_operator: AssertionOperator | None = None, assertion_expected: Any = None, message: str | None = None) -> float:
    return verify_assertion(249.99, assertion_operator, assertion_expected, "Price", message)

@keyword
def get_count(self, assertion_operator: AssertionOperator | None = None, assertion_expected: Any = None, message: str | None = None) -> int:
    return int_str_verify_assertion(3, assertion_operator, assertion_expected, "Count", message)

EOF cat > probe.robot <<'EOF' *** Settings *** Library Probe.py http://x

*** Test Cases *** Conversion ${a}= Union Float Str 249.99 Log To Console \nunion float|str "249.99" -> ${a} ${b}= Union Float Str value > 1 Log To Console union float|str "value > 1" -> ${b} ${c}= Any Expected 249.99 Log To Console Any "249.99" -> ${c} Helper Float Equal ${p}= Get Price == 249.99 Log To Console \nfloat_str == 249.99 ok -> ${p} Get Price > 100 Get Price validate value > 100 ${t}= Get Price then value * 2 Log To Console then -> ${t} Plain Verify Float Equal Get Price Plain == 249.99 Helper Float Wrong Get Price == 10 message=Custom {value} vs {expected} Helper Contains Rejected Get Price contains 9 Count Helper Get Count == 3 Get Count >= 1 Count Helper Wrong Get Count == 4 EOF ../demoshop-library/.venv/bin/python -m robot --output NONE --report NONE --log NONE --console verbose probe.robot 2>&1 | tail -40`*

/bin/bash: line 125: ../demoshop-library/.venv/bin/python: No such file or directory
Shell cwd was reset to <repo>

The agent runs cd <scratch> && ../rh/demoshop-library/.venv/bin/python -m robot --output NONE --report NONE --log NONE --console verbose probe.robot 2>&1 | tail -40

==============================================================================
Probe                                                                         
==============================================================================
Conversion                                                            
union float|str "249.99" -> str:'249.99'
union float|str "value > 1" -> str:'value > 1'
Any "249.99" -> str:'249.99'
| PASS |
------------------------------------------------------------------------------
Helper Float Equal                                                    
float_str == 249.99 ok -> 249.99
then -> 499.98
... (24 more lines)

<!DOCTYPE html>
<html lang="en">

<head>
  <meta charset="utf-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1" />
  <title>Flowline Supply | Shop the Future</title>
  <link rel="stylesheet" href="/assets/styles.bcca56d0f15fa943.css" />
  <script src="/static/app.js?v=ee3c3f9e45a296f9" defer></script>
  

<script type="application/ld+json">
... (41 more lines)

app.js
 const SESSION_COOKIE = "session_id";
 const SESSION_COOKIE_DAYS = 30;
 let sessionId = ensureSessionId();
 function ensureSessionId() {
   let existing = getCookie(SESSION_COOKIE);
     setCookie(SESSION_COOKIE, existing, SESSION_COOKIE_DAYS);
 // Cart badge and add to cart
 function getCartBadge() {
 function updateCartBadge(state) {
   const counter = getCartBadge();
   const quantity = items.reduce(
... (12 more lines)

Agent:

The shop's cart badge adds up item quantities (app.js:187), so Get Cart Item Count will do the same. One more check: Robot's TEST scope with PythonLibCore, including what a suite setup sees.

*The agent runs `cd && cat > ScopeProbe.py <<'EOF' import uuid from robotlibcore import DynamicCore, keyword

class Cart: def init(self, ctx): self.ctx = ctx @keyword def session_of_instance(self): return self.ctx.session_id

class ScopeProbe(DynamicCore): ROBOT_LIBRARY_SCOPE = "TEST" def init(self, url: str, space: str | None = None): self.session_id = uuid.uuid4().hex[:8] DynamicCore.init(self, [Cart(self)]) EOF cat > scope.robot <<'EOF' *** Settings *** Library ScopeProbe.py http://x Library ScopeProbe.py http://x space=other AS Other Suite Setup Log Session suite setup

*** Keywords *** Log Session [Arguments] ${where} ${s}= ScopeProbe.Session Of Instance Log To Console \n${where}: ${s}

*** Test Cases *** T1 [Setup] Log Session t1 setup Log Session t1 body ${o}= Other.Session Of Instance Log To Console t1 Other: ${o} T2 Log Session t2 body EOF ../rh/demoshop-library/.venv/bin/python -m robot --output NONE --report NONE --log NONE scope.robot 2>&1 | grep -E ':|PASS|FAIL' | head -20`*

suite setup: fe7a98ce
t1 setup: f80bc06d
t1 body: f80bc06d
t1 Other: de57215f
| PASS |
t2 body: 6712febb
| PASS |
Scope                                                                 | PASS |
Output:  NONE
Shell cwd was reset to <repo>

The agent runs openspec new change "add-demoshop-library" && openspec status --change "add-demoshop-library" --json

- Creating change 'add-demoshop-library' with schema 'spec-driven'...
Created change 'add-demoshop-library' at openspec/changes/add-demoshop-library/
Schema: spec-driven
Next: openspec status --change add-demoshop-library
{
  "changeName": "add-demoshop-library",
  "schemaName": "spec-driven",
  "planningHome": {
    "kind": "repo",
    "root": "<repo>",
    "changesDir": "<repo>/openspec/changes",
    "defaultSchema": "spec-driven"
... (98 more lines)

The agent runs openspec instructions proposal --change "add-demoshop-library" --json | jq 'del(.context)'

{
  "changeName": "add-demoshop-library",
  "artifactId": "proposal",
  "schemaName": "spec-driven",
  "changeDir": "<repo>/openspec/changes/add-demoshop-library",
  "planningHome": {
    "kind": "repo",
    "root": "<repo>",
    "changesDir": "<repo>/openspec/changes",
    "defaultSchema": "spec-driven"
  },
  "outputPath": "proposal.md",
... (15 more lines)

The agent runs openspec list --specs; openspec instructions proposal --change "add-demoshop-library" --json | jq '{rules, ctx_len: (.context|length)}'

No specs found.
{
  "rules": null,
  "ctx_len": 1428
}

Agent:

No specs exist yet, so every capability is new. Writing the proposal.

The agent writes <repo>/openspec/changes/add-demoshop-library/proposal.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent runs openspec status --change "add-demoshop-library" --json | jq -c '.artifacts[] | {id,status}'; openspec instructions specs --change "add-demoshop-library" --json | jq 'del(.context)'

{"id":"proposal","status":"done"}
{"id":"specs","status":"ready"}
{"id":"design","status":"ready"}
{"id":"tasks","status":"blocked"}
{
  "changeName": "add-demoshop-library",
  "artifactId": "specs",
  "schemaName": "spec-driven",
  "changeDir": "<repo>/openspec/changes/add-demoshop-library",
  "planningHome": {
    "kind": "repo",
    "root": "<repo>",
... (25 more lines)

Agent:

Now the four capability specs.

The agent writes <repo>/openspec/changes/add-demoshop-library/specs/library-import/spec.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent writes <repo>/openspec/changes/add-demoshop-library/specs/getter-assertions/spec.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent writes <repo>/openspec/changes/add-demoshop-library/specs/product-keywords/spec.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent writes <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/spec.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

Agent:

Before writing the design, I'll check the exact failure messages quoted in the specs against the real AssertionEngine and Robot behaviour, in the scratch directory.

*The agent runs `cd && cat > Msg.py <<'EOF' from typing import Any from assertionengine import AssertionOperator, float_str_verify_assertion, int_str_verify_assertion from robotlibcore import DynamicCore, keyword

class Msg(DynamicCore): def init(self): DynamicCore.init(self, []) @keyword def price_of(self, product_id: int, assertion_operator: AssertionOperator | None = None, assertion_expected: float | str = 0, message: str | None = None) -> float: return float_str_verify_assertion(249.99, assertion_operator, assertion_expected, f"Product {product_id} price", message) @keyword def total(self, assertion_operator: AssertionOperator | None = None, assertion_expected: float | str = 0, message: str | None = None) -> float: return float_str_verify_assertion(float(0), assertion_operator, assertion_expected, "Cart total", message) @keyword def items(self, assertion_operator: AssertionOperator | None = None, assertion_expected: int | str = 0, message: str | None = None) -> int: return int_str_verify_assertion(0, assertion_operator, assertion_expected, "Cart item count", message) EOF cat > msg.robot <<'EOF' *** Settings *** Library Msg.py

*** Test Cases *** Price Mismatch Price Of 1 == 10 Count Mismatch Items > 0 Custom Total == 5 message=Total was {value}, wanted {expected} Validate Fails Price Of 1 validate value < 100 Validate Passes And Alias And Named Price Of 1 validate 100 < value < 300 Price Of 1 greater than 0 Total assertion_operator=>= assertion_expected=0 Price Of 1 should be 249.99 Bad Operator Items === 12 Bad Id Price Of abc Matches Rejected Price Of 1 matches 249 EOF ../rh/demoshop-library/.venv/bin/python -m robot --output NONE --report NONE --log NONE msg.robot 2>&1 | grep -vE '^-+$|^=+$'`*

Msg                                                                           
Price Mismatch                                                        | FAIL |
Product 1 price '249.99' (float) should be '10.0' (float)
Count Mismatch                                                        | FAIL |
Cart item count '0' (int) should be greater than '0' (int)
Custom                                                                | FAIL |
Total was 0.0, wanted 5.0
Validate Fails                                                        | FAIL |
Product 1 price '249.99' (float) should validate to true with 'value < 100' (str)
Validate Passes And Alias And Named                                   | PASS |
Bad Operator                                                          | FAIL |
ValueError: Argument 'assertion_operator' got value '===' that cannot be converted to AssertionOperator or None.
... (8 more lines)

Agent:

All the quoted messages match exactly. Next, the design.

The agent runs openspec instructions design --change "add-demoshop-library" --json | jq '{instruction, template, rules, dependencies}'

{
  "instruction": "Create the design document that explains HOW to implement the change.\n\nWhen to include design.md (create only if any apply):\n- Cross-cutting change (multiple services/modules) or 
  "template": "# Design\n\n## Context\n\n<!-- Current state and constraints that shape the approach. See proposal.md for motivation - don't restate it -->\n\n## Goals / Non-Goals\n\n**Goals:**\n<!-- W
  "rules": null,
  "dependencies": [
    {
      "id": "proposal",
      "done": true,
      "path": "proposal.md",
      "description": "Initial proposal document outlining the change"
    }
  ]
... (1 more lines)

The agent reads <repo>/openspec/changes/add-demoshop-library/proposal.md

The agent writes <repo>/openspec/changes/add-demoshop-library/design.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent runs openspec status --change "add-demoshop-library" --json | jq -c '.artifacts[] | {id,status}'; openspec instructions tasks --change "add-demoshop-library" --json | jq '{instruction, template, rules, dependencies}'

{"id":"proposal","status":"done"}
{"id":"specs","status":"done"}
{"id":"design","status":"done"}
{"id":"tasks","status":"ready"}
{
  "instruction": "Create the task list that breaks down the implementation work.\n\nBefore writing tasks, check design.md for Open Questions. If any of them\nwould change what gets built, resolve them
  "template": "# Tasks\n\n## 1. <!-- Task Group Name -->\n\n- [ ] 1.1 <!-- Task description -->\n- [ ] 1.2 <!-- Task description -->\n\n## 2. <!-- Task Group Name -->\n\n- [ ] 2.1 <!-- Task descriptio
  "rules": null,
  "dependencies": [
    {
      "id": "specs",
      "done": true,
... (11 more lines)

Agent:

Writing the tasks. I'll also add the README to the proposal's Impact section, since one task fills in the empty README and the placeholder package description.

The agent writes <repo>/openspec/changes/add-demoshop-library/tasks.md

File created successfully at: <repo>/openspec/changes/add-demoshop-library/

The agent edits <repo>/openspec/changes/add-demoshop-library/proposal.md

The file <repo>/openspec/changes/add-demoshop-library/proposal.md has been

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent runs openspec validate add-demoshop-library --strict; echo "---"; openspec status --change "add-demoshop-library"; echo "---"; find openspec/changes/add-demoshop-library -type f | sort; git status --short; rm -rf ../../rf-probe

Change 'add-demoshop-library' is valid
---
- Loading change status...
Change: add-demoshop-library
Schema: spec-driven
Change root: <repo>/openspec/changes/add-demoshop-library
Progress: 4/4 artifacts complete

[x] proposal
[x] specs
[x] design
[x] tasks
... (22 more lines)

Agent:

I created the change add-demoshop-library in openspec/changes/add-demoshop-library/. It passes openspec validate --strict. No project code was changed.

What I created:

proposal.md: why the library is needed, the five keywords, and four new capabilities. It also lists what the change touches: shop endpoints, tests/, atest/, docs and README.
Specs, one per capability, with every scenario written so it can be tested:
- library-import: importing with url and space, the X-Workshop-Space header, one cart per instance, the library scope, error reporting and libdoc.
- getter-assertions: the shared assertion_operator / assertion_expected / message contract and the exact failure messages.
- product-keywords: Get Product Count and Get Product Price.
- cart-keywords: Add Product To Cart, Get Cart Item Count and Get Cart Total.
design.md: nine decisions, each with the alternatives considered, plus the response shapes I saw on the local shop. The OpenAPI file gives no response schemas.
tasks.md: 20 tasks in 8 groups, each saying how to check it is done.

I made three decisions you should review, because each changes what tests will see:

Each test gets a fresh library instance (Robot's TEST scope), so every test starts with an empty cart. This slice has no Clear Cart keyword, so if instances lived for a whole suite, each cart test would depend on the ones before it. The catch is that a cart filled in a suite setup is not visible to the tests.
Get Cart Item Count adds up quantities, so two of product 1 plus one of product 2 counts as 3, not 2. That is what the shop's own cart badge shows (/static/app.js).
The Get keywords check with AssertionEngine's numeric helpers (int_str_verify_assertion and float_str_verify_assertion), not by calling verify_assertion directly. Robot 7.5 passes "249.99" in as text, and a direct call then fails with '249.99' (float) should be '249.99' (str). The helpers convert the expected value first and then call verify_assertion, so Get Product Price 1 == 249.99 passes.

Things I found and checked:

Session header: without an X-Session-ID header, every caller shares the shop's workshop-demo cart. So the library generates its own session ID per instance and sends it with every request.
Assertion messages: I tried every assertion message quoted in the specs against the installed Robot Framework, AssertionEngine and PythonLibCore in a scratch directory, which I have since deleted.
Probe of the shop: I added items to a throwaway cart to see the cart's shape and the 404 and 422 responses, then cleared it. Inventory was unchanged.

Still open: whether the generated docs/DemoShopLibrary.html gets committed or ignored. Task 6.2 settles it.

The artifacts are ready for review. When you are ready, run /opsx:apply.

Step 6 - Review the proposal

The rehearsal's review: the questions of step 6, put to the agent, which changes its artifacts where an answer is no.

Prompt:

Review the change you just proposed against these questions, and change its artifacts wherever an answer is no. Then answer each question in one line.

Is every keyword named and argued like the examples: Get <thing>, then the three assertion arguments? Does every check go through AssertionEngine, with no assertion code of the library's own? Do the endpoints, parameters and headers match references/demoshop-openapi.json? Do the Robot tests in atest/ use only the library's keywords, against the local shop?

Agent:

I'll re-read the artifacts from disk and check them against the four questions.

The agent runs cd <repo>/openspec/changes/add-demoshop-library && grep -nE 'X-Session-ID|X-Workshop-Space|Accept|Evaluate|Run Keyword|Should |localhost:1|AS +|every request|stay as it was|unchanged|from before' proposal.md design.md tasks.md specs/*/spec.md

proposal.md:14:  optional `space`. When `space` is given, every request carries it in the `X-Workshop-Space` header.
proposal.md:16:  that send no `X-Session-ID`, so the library generates a session id per instance and sends it with every request.
design.md:21:- A request without `X-Session-ID` uses the shared cart `"session": "workshop-demo"`.
design.md:74:`Accept: application/json` and `X-Session-ID: <session_id>`, and `X-Workshop-Space: <space>` only when `space` is not
design.md:151:requests: path, `X-Session-ID`, `X-Workshop-Space` and body.
design.md:163:### 8. Acceptance tests: one resource and one suite per capability
design.md:176:single space. Expected failures use `Run Keyword And Expect Error`. The check that the default cart is left untouched
design.md:178:`Evaluate    requests.get(...).json()    modules=requests`, before and after adding a product. The suites run with
tasks.md:17:  `session_id`, and creates a `requests.Session` with the default headers `Accept`, `X-Session-ID` and, only when
tasks.md:18:  `space` is set, `X-Workshop-Space`. It also provides `request(method, path, action, json=None)` with a 10-second
tasks.md:20:  `/`; the space header is present only when `space` is given; two clients have different session ids; every request
tasks.md:21:  carries `X-Session-ID`; a POST body is sent as JSON.
... (22 more lines)

The agent writes <repo>/openspec/changes/add-demoshop-library/specs/library-import/spec.md

The file <repo>/openspec/changes/add-demoshop-library/specs/library-import/

The agent edits <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/spec.md

The file <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/s

The agent edits <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/spec.md

The file <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/s

Agent:

Next, the proposal.

The agent edits <repo>/openspec/changes/add-demoshop-library/proposal.md

The file <repo>/openspec/changes/add-demoshop-library/proposal.md has been

The agent edits <repo>/openspec/changes/add-demoshop-library/proposal.md

The file <repo>/openspec/changes/add-demoshop-library/proposal.md has been

The agent edits <repo>/openspec/changes/add-demoshop-library/proposal.md

The file <repo>/openspec/changes/add-demoshop-library/proposal.md has been

Agent:

Now the design. I'll re-read it before editing the sections that change.

The agent reads <repo>/openspec/changes/add-demoshop-library/design.md

Agent:

Before writing the acceptance-test check into the design, I'll confirm that Robot 7.5's result model exposes which library a keyword came from.

The agent runs <repo>/.venv/bin/python -c " from robot.result import Keyword k = Keyword(name='Get Cart Total', owner='ShopA') print(k.owner, '|', k.full_name) "

ShopA | ShopA.Get Cart Total

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

Agent:

Now the tasks.

The agent reads <repo>/openspec/changes/add-demoshop-library/tasks.md

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent edits <repo>/openspec/changes/add-demoshop-library/design.md

The file <repo>/openspec/changes/add-demoshop-library/design.md has been up

The agent runs cd <repo> && openspec validate add-demoshop-library --strict && openspec status --change add-demoshop-library && echo "--- leftovers:" && grep -nE 'Accept|Run Keyword And Expect|Evaluate +requests|localhost:1|OtherShop|stay as it was|every request\b' openspec/changes/add-demoshop-library/{proposal.md,design.md,tasks.md} openspec/changes/add-demoshop-library/specs/*/spec.md

Change 'add-demoshop-library' is valid
- Loading change status...
Change: add-demoshop-library
Schema: spec-driven
Change root: <repo>/openspec/changes/add-demoshop-library
Progress: 4/4 artifacts complete

[x] proposal
[x] specs
[x] design
[x] tasks

... (17 more lines)

Agent:

Keywords like the examples? Yes, no change needed: the four Get keywords take the keyword's own argument (product_id for Get Product Price), then assertion_operator, assertion_expected and message, typed like Browser's Get Element Count; only Add Product To Cart, the name you asked for, is not a getter, so it has no assertion arguments.
Every check through AssertionEngine? Yes, no change needed: every check calls AssertionEngine's int_str_verify_assertion or float_str_verify_assertion, which call verify_assertion, and the library itself only translates shop errors.
Endpoints, parameters and headers match the OpenAPI document? No, now fixed: X-Session-ID goes only on cart requests, where the document declares it, and the undeclared Accept header is gone; X-Workshop-Space is still undeclared but stays because AGENTS.md requires it, which the design now says.
atest/ uses only the library's keywords, against the local shop? No, now fixed: I removed Evaluate with requests, Run Keyword And Expect Error, the localhost:1 test and the ambiguous two-instance import (now ShopA/ShopB), so the suites call only library keywords against ${SHOP_URL}, failure cases move to the unit tests, and new task 7.6 checks this in output.xml.

I also dropped "the cart stays as it was" from the cart spec, because that is the shop's behaviour and no test could check it. The change still passes openspec validate --strict.

Step 7 - Apply it

The rehearsal tool stopped the first /opsx:apply session after 13 of the change's 22 tasks, at its own time limit; that session is not recorded. A new /opsx:apply continued from the ticked tasks, as you would after an interruption.

Prompt:

/opsx:apply

The agent runs openspec list --json

{
  "changes": [
    {
      "name": "add-demoshop-library",
      "completedTasks": 13,
      "totalTasks": 22,
      "lastModified": "2026-10-06T18:26:40.002Z",
      "status": "in-progress"
    }
  ],
  "root": {
    "path": "<repo>",
... (3 more lines)

Agent:

Using change: add-demoshop-library (override with /opsx:apply <other>).

The agent runs openspec status --change "add-demoshop-library" --json && echo ---- && openspec instructions apply --change "add-demoshop-library" --json

{
  "changeName": "add-demoshop-library",
  "schemaName": "spec-driven",
  "planningHome": {
    "kind": "repo",
    "root": "<repo>",
    "changesDir": "<repo>/openspec/changes",
    "defaultSchema": "spec-driven"
  },
  "changeRoot": "<repo>/openspec/changes/add-demoshop-library",
  "artifactPaths": {
    "proposal": {
... (242 more lines)

The agent reads <repo>/openspec/changes/add-demoshop-library/tasks.md

The agent reads <repo>/openspec/changes/add-demoshop-library/design.md

The agent reads <repo>/openspec/changes/add-demoshop-library/proposal.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/getter-assertions/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/library-import/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/product-keywords/spec.md

Agent:

Now the existing code and references.

The agent runs git ls-files --others --exclude-standard | grep -v '^openspec/' ; echo ---; ls -la atest docs 2>&1; cat .gitignore; echo ---; cat pyproject.toml; echo ---; cat README.md

.claude/commands/opsx/apply.md
.claude/commands/opsx/archive.md
.claude/commands/opsx/explore.md
.claude/commands/opsx/propose.md
.claude/commands/opsx/sync.md
.claude/commands/opsx/update.md
.claude/skills/openspec-apply-change/SKILL.md
.claude/skills/openspec-archive-change/SKILL.md
.claude/skills/openspec-explore/SKILL.md
.claude/skills/openspec-propose/SKILL.md
.claude/skills/openspec-sync-specs/SKILL.md
.claude/skills/openspec-update-change/SKILL.md
... (90 more lines)

The agent reads <repo>/src/demoshop_library/library.py

The agent reads <repo>/src/demoshop_library/products.py

The agent reads <repo>/src/demoshop_library/cart.py

The agent reads <repo>/references/browser-get-text.md

The agent reads <repo>/references/browser-get-element-count.md

Agent:

Keyword docs and atest/ files already exist even though their tasks are unchecked. Let me look at them and the remaining source.

The agent runs for f in atest/resources/demoshop.resource atest/*.robot src/demoshop_library/__init__.py src/demoshop_library/assertions.py src/demoshop_library/client.py; do echo "=== $f"; cat -A "$f" | grep -c '\^M' ; cat "$f"; done

=== atest/resources/demoshop.resource
0
*** Settings ***
Documentation       Variables shared by the acceptance suites: where the shop runs and its seed data.
...                 Override any of them with ``--variable``, for example ``--variable WORKSHOP_SPACE:team-7``.


*** Variables ***
${SHOP_URL}             http://localhost:9090
# Not ${SPACE}, which is Robot Framework's built-in variable holding a single space.
${WORKSHOP_SPACE}       ${NONE}

... (270 more lines)

The agent runs uv run pytest -q 2>&1 | tail -5; echo ---; curl -s -m 5 -o /dev/null -w '%{http_code}\n' http://localhost:9090/api/products/ ; curl -s -m 5 http://localhost:9090/health; echo

........................................................................ [ 94%]
....                                                                     [100%]
76 passed in 0.17s
---
200
{"status":"ok","version":"0.3.0"}

The agent runs uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html && uv run python -m robot.libdoc --specdocformat HTML demoshop_library.DemoShopLibrary <scratch> >/dev/null && uv run python - <<'EOF' import json, re d = json.load(open("<scratch>")) print("intro has Assertions anchor:", 'id="Assertions"' in d["doc"]) print("intro anchors:", re.findall(r'<h2 id="([^"]+)"', d["doc"])) print("init doc names url/space:", all(f"<code>{a}</code>" in d["inits"][0]["doc"] for a in ("url", "space"))) for kw in d["keywords"]: doc = kw["doc"] links = set(re.findall(r'href="#([^"]+)"', doc)) args = [a["name"] for a in kw["args"]] print(f'- {kw["name"]}: args={args} tags={kw["tags"]} links={sorted(links)} ' f'table={"=Arguments=" in doc or "<th>Arguments</th>" in doc} example={"Example:" in doc} ' f'optional={"Optionally asserts" in doc}') for l in links: assert l in ("Assertions", "Carts and library scope") or l.replace("%20"," ") in [k["name"] for k in d["keywords"]], l EOF grep -o 'id="Carts and library scope"\|id="Assertions"' /dev/null; ls -la docs/

Library 'demoshop_library.DemoShopLibrary' expected 1 to 2 arguments, got 0.

Try --help for usage information.
total 236
drwxr-xr-x  2 user user   4096 Oct  6 20:26 .
drwxr-xr-x 13 user user   4096 Oct  6 20:27 ..
-rw-r--r--  1 user user 230311 Oct  6 20:26 DemoShopLibrary.html

Agent:

libdoc fails with the exact command that task 6.1, the library-import spec and AGENTS.md all prescribe. I'll check how Robot 7.5's libdoc builds a library before deciding what to do.

       name, args = self._split_library_name_and_args(library)
       lib = TestLibrary.from_name(name, args=args)
   def _split_library_name_and_args(self, library):
       args = library.split(self._argument_separator)
       name = args.pop(0)
       return self._normalize_library_path(name), args
       if lib.init.args:
           for name, type_info in self._yield_names_and_infos(kw.args):
   def _yield_names_and_infos(self, args: ArgumentSpec):
       for arg in args:
       if args.return_type:
           for type_info in self._yield_infos(args.return_type):
... (28 more lines)

The agent runs uv run python -m robot.libdoc demoshop_library.DemoShopLibrary::http://localhost:9090 docs/DemoShopLibrary.html && uv run python -m robot.libdoc --specdocformat HTML demoshop_library.DemoShopLibrary::http://localhost:9090 <scratch> >/dev/null && uv run python - <<'EOF' import json, re d = json.load(open("<scratch>")) anchors = re.findall(r'<h2 id="([^"]+)"', d["doc"]) print("intro sections:", anchors) print("init args:", [a["name"] for a in d["inits"][0]["args"]]) names = [k["name"] for k in d["keywords"]] for kw in d["keywords"]: doc = kw["doc"] links = sorted(set(re.findall(r'href="#([^"]+)"', doc))) bad = [l for l in links if l not in anchors and l.replace("%20", " ") not in names] print(f'- {kw["name"]}({", ".join(a["repr"] for a in kw["args"])}) tags={kw["tags"]}') print(f' links={links} broken={bad} optional_asserts={"Optionally asserts" in doc} example={"Example:" in doc}') if kw["name"].startswith("Get"): row = re.search(r'<code>assertion_operator</code></td>\s*<td>(.*?)</td>', doc).group(1) print(" operator row:", re.sub(r"<[^>]+>", "", row), "| links Assertions:", 'href="#Assertions"' in row) EOF

<repo>/docs/DemoShopLibrary.html
intro sections: ['Shop and workshop space', 'Carts and library scope', 'Assertions', 'Shop errors']
init args: ['url', 'space']
- Add Product To Cart(product_id: int, quantity: int = 1) tags=['Cart', 'Setter']
    links=['Add%20Product%20To%20Cart', 'Carts%20and%20library%20scope', 'Get%20Cart%20Item%20Count'] broken=['Carts%20and%20library%20scope'] optional_asserts=False example=True
- Get Cart Item Count(assertion_operator: AssertionOperator | None = None, assertion_expected: int | str = 0, message: str | None = None) tags=['Assertion', 'Cart', 'Getter']
    links=['Add%20Product%20To%20Cart', 'Assertions', 'Get%20Cart%20Item%20Count'] broken=[] optional_asserts=True example=True
    operator row: See Assertions for further details. Defaults to None. | links Assertions: True
- Get Cart Total(assertion_operator: AssertionOperator | None = None, assertion_expected: float | str = 0, message: str | None = None) tags=['Assertion', 'Cart', 'Getter']
    links=['Add%20Product%20To%20Cart', 'Assertions', 'Get%20Cart%20Total'] broken=[] optional_asserts=True example=True
    operator row: See Assertions for further details. Defaults to None. | links Assertions: True
- Get Product Count(assertion_operator: AssertionOperator | None = None, assertion_expected: int | str = 0, message: str | None = None) tags=['Assertion', 'Getter', 'Products']
... (5 more lines)

The agent runs uv run robot --dryrun --outputdir <scratch> atest 2>&1 | tail -25

------------------------------------------------------------------------------
Atest.Library Import :: Importing the library: URL, workshop space... | PASS |
6 tests, 6 passed, 0 failed
==============================================================================
Atest.Product Keywords :: Product keywords against the shop's seed catalogue. 
==============================================================================
Count Of The Seed Catalogue                                           | PASS |
------------------------------------------------------------------------------
Count With Assertion                                                  | PASS |
------------------------------------------------------------------------------
Price Of A Product                                                    | PASS |
------------------------------------------------------------------------------
... (13 more lines)

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

Agent:

✓ Task 7.1 complete (dry run: 26 tests, 0 failures).

Working on tasks 7.2–7.5: each suite exists and covers the scenarios its task lists. Next I'll run each one against the shop.

##### product_keywords
5 tests, 5 passed, 0 failed
rc=0
##### cart_keywords
7 tests, 7 passed, 0 failed
rc=0
##### getter_assertions
8 tests, 8 passed, 0 failed
rc=0
##### library_import
6 tests, 6 passed, 0 failed
rc=0

Agent:

All four suites pass against the shop: 5 + 7 + 8 + 6 tests, with no errors or warnings.

The agent runs sed -i -E 's/^- \[ \] (7\.[2-5] )/- [x] \1/' openspec/changes/add-demoshop-library/tasks.md && grep -n '^- \[' openspec/changes/add-demoshop-library/tasks.md | sed -n '13,22p'

- [ ] 6.1 Write the documentation of the five keywords in the shape of `references/browser-get-text.md` and
- [x] 6.2 Settle the open question about generated docs: commit `docs/DemoShopLibrary.html`, or add `docs/` to
- [x] 7.1 Create `atest/resources/demoshop.resource` holding variables only: `${SHOP_URL}` (default
- [x] 7.2 Write `atest/product_keywords.robot`. It imports the library with `url=${SHOP_URL}` and
- [x] 7.3 Write `atest/cart_keywords.robot`. It covers the passing scenarios of `specs/cart-keywords/spec.md`:
- [x] 7.4 Write `atest/getter_assertions.robot`. It covers the passing scenarios of `specs/getter-assertions/spec.md`:
- [x] 7.5 Write `atest/library_import.robot`. It imports the library twice, as `ShopA` with `url=${SHOP_URL}` and as
- [ ] 7.6 Check that the acceptance tests use only the library's keywords. Verify by reading `results/output.xml`
- [ ] 8.1 Add a README usage section with the import example, the five keywords, and the commands for unit tests,
- [ ] 8.2 Run the whole suite. Verify that `uv run pytest` and `uv run robot --outputdir results atest` both pass with

Agent:

✓ Tasks 7.2–7.5 complete.

Working on task 7.6: full run, then I'll check every executed keyword's owner in results/output.xml.

*The agent runs `rm -rf results/product_keywords results/cart_keywords results/getter_assertions results/library_import && uv run robot --outputdir results --console quiet atest; echo "robot rc=$?"; uv run python - <<'EOF' from collections import Counter from robot.api import ExecutionResult, ResultVisitor

ALLOWED = {"demoshop_library.DemoShopLibrary", "ShopA", "ShopB"}

class Owners(ResultVisitor): def init(self): self.owners, self.bad, self.other = Counter(), [], Counter() def start_keyword(self, kw): self.owners[kw.owner] += 1 if kw.owner not in ALLOWED: self.bad.append((kw.parent.name if kw.parent else "?", kw.full_name)) def start_body_item(self, item): if item.type not in ("KEYWORD", "SETUP", "TEARDOWN"): self.other[item.type] += 1 super().start_body_item(item)

result = ExecutionResult("results/output.xml") v = Owners() result.visit(v) s = result.statistics.total print(f"tests: {s.total} passed={s.passed} failed={s.failed}") print("owners:", dict(v.owners)) print("non-keyword body items:", dict(v.other)) print("violations:", v.bad) assert not v.bad and s.failed == 0 print("7.6 check: OK") EOF`*

robot rc=0
tests: 26 passed=26 failed=0
owners: {'demoshop_library.DemoShopLibrary': 54, 'ShopA': 9, 'ShopB': 5}
non-keyword body items: {'MESSAGE': 6}
violations: []
7.6 check: OK

Agent:

All 26 tests pass. The 68 executed keywords are owned only by demoshop_library.DemoShopLibrary (54), ShopA (9) and ShopB (5). The 6 MESSAGE items are log lines, not keywords.

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

Agent:

✓ Task 7.6 complete.

Before pausing, I'll run the one 8.2 check I haven't run yet, the wheel contents. I'm building into a temp directory so nothing is left in the repo.

The agent runs uv build --out-dir <scratch> 2>&1 | tail -2 && uv run python -c " import zipfile, glob names = zipfile.ZipFile(glob.glob('<scratch>')[0]).namelist() print('\n'.join(names)) print('tests/atest in wheel:', [n for n in names if n.startswith(('tests/', 'atest/'))]) "

Successfully built <scratch>
Successfully built <scratch>
demoshop_library/
demoshop_library/__init__.py
demoshop_library/assertions.py
demoshop_library/cart.py
demoshop_library/client.py
demoshop_library/library.py
demoshop_library/products.py
demoshop_library/py.typed
demoshop_library-0.1.0.dist-info/
demoshop_library-0.1.0.dist-info/WHEEL
... (3 more lines)

Agent:

Implementation paused

Change: add-demoshop-library (spec-driven) Progress: 19 of 22 tasks done

Done this session

Most of the remaining work was already on disk but had never been checked. I ran each task's check against the shop at localhost:9090 (version 0.3.0) and only marked tasks whose check passed:

7.1 demoshop.resource: the dry run has 26 tests, all passing.
7.2–7.5 Product (5 tests), cart (7), assertion (8) and import (6) suites: all pass against the shop.
7.6 Full run: 26 of 26 pass. Every executed keyword belongs to demoshop_library.DemoShopLibrary, ShopA or ShopB.

Problem: libdoc fails with the command the plan prescribes

Task 6.1, the library-import spec ("Generate documentation") and AGENTS.md all use:

uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html

It fails with Library 'demoshop_library.DemoShopLibrary' expected 1 to 2 arguments, got 0. libdoc has to create the library to list its keywords, and the spec makes url a required argument. The plan contradicts itself here; this is not a coding mistake.

The documentation itself is finished. With the URL passed in, libdoc succeeds, lists all five keywords, and every Get keyword's assertion_operator row links to Assertions. So 6.1 stays unchecked only because its check fails as written.

Options

Pass the URL to libdoc (recommended): uv run python -m robot.libdoc demoshop_library.DemoShopLibrary::http://localhost:9090 docs/DemoShopLibrary.html. This works now with no code change. Importing never contacts the shop, so any URL would work. The command needs updating in AGENTS.md, the spec scenario and task 6.1 (/opsx:update covers the plan files).
Make url optional, defaulting to http://localhost:9090. The plain command would then work, but this changes the spec's "required url" rule, the design and the unit tests.

Still to do once you decide

6.1: mark it done once its check matches your choice.
8.1: write the README usage section. It includes the libdoc command, so I held off.
8.2: I ran its checks now and they all pass: 76 unit tests, 26 acceptance tests, the 7.6 check, and a wheel with the required modules and py.typed but no tests/ or atest/. I'll mark it done only after a final run, because option 2 changes code.

Which option do you want? Then run /opsx:apply again to finish.

Step 7 - Apply it, continued

The apply paused on a defect of the lab's first text: its libdoc command did not pass the library's url. The lab now gives the corrected command. The rehearsal told the agent, as you would, and applied the rest.

Prompt:

Your apply paused on the libdoc command: libdoc imports the library, so it needs the url the library takes. Use this command in AGENTS.md and in the change's tasks instead: uv run python -m robot.libdoc "demoshop_library.DemoShopLibrary::url=http://localhost:9090" docs/DemoShopLibrary.html. Change nothing else.

The agent searches for robot\.libdoc

AGENTS.md:11:  `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html`.
openspec/config.yaml:13:  - Keyword documentation: `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html`.
openspec/changes/add-demoshop-library/tasks.md:99:  `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html` succeeds, and that the
openspec/changes/add-demoshop-library/specs/library-import/spec.md:108:- **WHEN** `uv run python -m robot.libdoc demoshop_library.DemoShopLibrary docs/DemoShopLibrary.html` is run

The agent reads <repo>/AGENTS.md

The agent reads <repo>/openspec/changes/add-demoshop-library/tasks.md

The agent edits <repo>/AGENTS.md

The file <repo>/AGENTS.md has been updated successfully.

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

Agent:

I replaced the libdoc command with yours in two places and changed nothing else:

AGENTS.md:11
openspec/changes/add-demoshop-library/tasks.md:99 (task 6.1)

I didn't run the command, because that would write docs/DemoShopLibrary.html.

The old command, without the url, is still in two places I left alone:

openspec/config.yaml:13
openspec/changes/add-demoshop-library/specs/library-import/spec.md:108, inside a scenario's WHEN step

The spec scenario will still disagree with task 6.1 until it's updated. Do you want me to change these two as well?

Prompt:

/opsx:apply

The agent runs openspec list --json

{
  "changes": [
    {
      "name": "add-demoshop-library",
      "completedTasks": 19,
      "totalTasks": 22,
      "lastModified": "2026-10-06T20:21:23.598Z",
      "status": "in-progress"
    }
  ],
  "root": {
    "path": "<repo>",
... (3 more lines)

Agent:

Using change: add-demoshop-library (override with /opsx:apply <other>).

The agent runs openspec status --change "add-demoshop-library" --json && openspec instructions apply --change "add-demoshop-library" --json

{
  "changeName": "add-demoshop-library",
  "schemaName": "spec-driven",
  "planningHome": {
    "kind": "repo",
    "root": "<repo>",
    "changesDir": "<repo>/openspec/changes",
    "defaultSchema": "spec-driven"
  },
  "changeRoot": "<repo>/openspec/changes/add-demoshop-library",
  "artifactPaths": {
    "proposal": {
... (241 more lines)

The agent reads <repo>/openspec/changes/add-demoshop-library/tasks.md

The agent reads <repo>/openspec/changes/add-demoshop-library/proposal.md

The agent reads <repo>/openspec/changes/add-demoshop-library/design.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/cart-keywords/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/getter-assertions/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/library-import/spec.md

The agent reads <repo>/openspec/changes/add-demoshop-library/specs/product-keywords/spec.md

The agent runs git status --short && ls -la && find src tests atest references docs -type f 2>/dev/null | grep -v __pycache__ && cat .gitignore pyproject.toml README.md

?? .claude/
?? .gitignore
?? .python-version
?? AGENTS.md
?? README.md
?? atest/
?? openspec/
?? pyproject.toml
?? references/
?? results-assert/
?? src/
?? tests/
... (90 more lines)

The agent reads <repo>/src/demoshop_library/library.py

The agent reads <repo>/src/demoshop_library/products.py

The agent reads <repo>/src/demoshop_library/cart.py

The agent reads <repo>/src/demoshop_library/assertions.py

The agent reads <repo>/src/demoshop_library/__init__.py

The agent reads <repo>/src/demoshop_library/client.py

The agent reads <repo>/references/browser-get-text.md

The agent reads <repo>/references/browser-get-element-count.md

The agent runs curl -s -m 5 http://localhost:9090/api/products/ | head -c 300; echo; echo "---"; uv run pytest -q 2>&1 | tail -5; ls results results-assert dist docs; git log --oneline 2>&1 | head

{"items":[{"id":5,"name":"Atlas Standing Desk","sku":"ATL-DSK-005","price":799.0,"description":"Programmable standing desk with posture coaching, ambient wellness reminders, and built-in cable managem
---
........................................................................ [ 94%]
....                                                                     [100%]
76 passed in 0.16s
dist:
demoshop_library-0.1.0-py3-none-any.whl
demoshop_library-0.1.0.tar.gz

docs:
DemoShopLibrary.html

... (10 more lines)

The agent runs uv run python -m robot.libdoc "demoshop_library.DemoShopLibrary::url=http://localhost:9090" docs/DemoShopLibrary.html; echo "exit=$?"; echo "--- without url arg:"; uv run python -m robot.libdoc demoshop_library.DemoShopLibrary /tmp/libdoc-nourl.html; echo "exit=$?"

<repo>/docs/DemoShopLibrary.html
exit=0
--- without url arg:
Library 'demoshop_library.DemoShopLibrary' expected 1 to 2 arguments, got 0.

Try --help for usage information.
exit=252

The agent runs uv run python - <<'EOF' import json, re html = open("docs/DemoShopLibrary.html", encoding="utf-8").read() m = re.search(r"libdoc = (\{.*?\});\s*</script>", html, re.S) data = json.loads(m.group(1)) intro = data["doc"] anchors = set(re.findall(r'<h2 id="([^"]+)"', intro)) | {"Importing", "Keywords"} print("intro sections:", sorted(re.findall(r'<h2 id="([^"]+)"', intro))) print("keywords:", [k["name"] for k in data["keywords"]]) for kw in [data["inits"][0] if data["inits"] else None] + data["keywords"]: if kw is None: continue links = re.findall(r'<a href="#([^"]+)" class="name">', kw["doc"]) broken = [l for l in links if l not in anchors and l not in {k["name"] for k in data["keywords"]}] print(f'{kw["name"]!r}: args={[a["repr"] for a in kw["args"]]} tags={kw.get("tags")} links={sorted(set(links))} broken={broken}') first = re.sub("<.*?>", "", kw["doc"]).strip().splitlines()[0] print(" summary:", first) print(" has Example:", "Example" in kw["doc"], "| optional-assert para:", "Optionally asserts" in kw["doc"]) intro_links = re.findall(r'<a href="#([^"]+)" class="name">', intro) print("intro links:", sorted(set(intro_links)), "broken:", [l for l in intro_links if l not in anchors and l not in {k["name"] for k in data["keywords"]}]) EOF

Exit code 1
Traceback (most recent call last):
  File "<stdin>", line 4, in <module>
AttributeError: 'NoneType' object has no attribute 'group'

The agent runs grep -o 'libdoc[^=]\{0,20\}=' docs/DemoShopLibrary.html | sort | uniq -c | head; grep -o '<script type="text/javascript">[^<]\{0,80\}' docs/DemoShopLibrary.html | head -5

libdoc =
libdoc"),rA=
libdoc-logo" aria-hidden=
libdoc-overview"><div id=
libdoc-title" href=
libdoc-title"),t=
libdoc.keywords.map(e=
libdoc.keywords].forEach(e=
libdoc.selectedTag=
libdoc;const rL=
<script type="text/javascript">

The agent runs uv run python - <<'EOF' import json, re html = open("docs/DemoShopLibrary.html", encoding="utf-8").read() i = html.index("libdoc = ") + len("libdoc = ") data, _ = json.JSONDecoder().raw_decode(html[i:]) intro = data["doc"] sections = re.findall(r'<h2 id="([^"]+)"', intro) kwnames = {k["name"] for k in data["keywords"]} anchors = set(sections) | {"Importing", "Keywords"} | kwnames print("intro sections:", sections) print("keywords:", [k["name"] for k in data["keywords"]]) for kw in data["inits"] + data["keywords"]: links = re.findall(r'<a href="#([^"]+)" class="name">', kw["doc"]) broken = [l for l in links if l not in anchors] print(f'{kw["name"]!r}: args={[a["repr"] for a in kw["args"]]} tags={kw.get("tags")}') print(" links:", sorted(set(links)), "broken:", broken) print(" summary:", re.sub("<.*?>", "", kw["doc"]).strip().splitlines()[0]) print(" Example:", "Example" in kw["doc"], "| Optionally asserts:", "Optionally asserts" in kw["doc"], "| table:", "Arguments" in kw["doc"]) intro_links = re.findall(r'<a href="#([^"]+)" class="name">', intro) print("intro links:", sorted(set(intro_links)), "broken:", [l for l in intro_links if l not in anchors]) EOF

intro sections: ['Shop and workshop space', 'Carts and library scope', 'Assertions', 'Shop errors']
keywords: ['Add Product To Cart', 'Get Cart Item Count', 'Get Cart Total', 'Get Product Count', 'Get Product Price']
'__init__': args=['url: str', 'space: str | None = None'] tags=[]
   links: [] broken: []
   summary: Imports the library for the shop at url.
   Example: True | Optionally asserts: False | table: True
'Add Product To Cart': args=['product_id: int', 'quantity: int = 1'] tags=['Cart', 'Setter']
   links: [] broken: []
   summary: Adds quantity of the product with product_id to the cart.
   Example: True | Optionally asserts: False | table: True
'Get Cart Item Count': args=['assertion_operator: AssertionOperator | None = None', 'assertion_expected: int | str = 0', 'message: str | None = None'] tags=['Assertion', 'Cart', 'Getter']
   links: [] broken: []
... (15 more lines)

The agent runs uv run python - <<'EOF' import json, re html = open("docs/DemoShopLibrary.html", encoding="utf-8").read() i = html.index("libdoc = ") + len("libdoc = ") data, _ = json.JSONDecoder().raw_decode(html[i:]) doc = next(k for k in data["keywords"] if k["name"] == "Get Product Price")["doc"] print(doc[:1500]) print("=== h2 sample:", re.findall(r'<h2[^>]*>', data["doc"])) EOF

<p>Returns the price of the product with <code>product_id</code> as a float.</p>
<table border="1">
<tr>
<th>Arguments</th>
<th>Description</th>
</tr>
<tr>
<td><code>product_id</code></td>
<td>Id of the product, an integer. An unknown id fails the keyword with the shop's <code>Product not found</code>.</td>
</tr>
<tr>
<td><code>assertion_operator</code></td>
... (19 more lines)

The agent runs uv run python - <<'EOF' import json, re from urllib.parse import unquote html = open("docs/DemoShopLibrary.html", encoding="utf-8").read() i = html.index("libdoc = ") + len("libdoc = ") data, _ = json.JSONDecoder().raw_decode(html[i:]) sections = set(re.findall(r'<h2 id="([^"]+)"', data["doc"])) anchors = sections | {"Importing", "Keywords"} | {k["name"] for k in data["keywords"]} for name, doc in [("__intro__", data["doc"])] + [(k["name"], k["doc"]) for k in data["inits"] + data["keywords"]]: links = {unquote(l) for l in re.findall(r'<a href="#([^"]+)"[^>]*class="name"', doc)} unresolved = re.findall(r'<span class="name">([^<]+)</span>', doc) print(f"{name}: links={sorted(links)} broken={sorted(links - anchors)} unresolved={unresolved}") EOF

__intro__: links=['Add Product To Cart', 'Assertions', 'Get Cart Item Count', 'Get Cart Total', 'Get Product Count', 'Get Product Price', 'Importing'] broken=[] unresolved=[]
__init__: links=['Carts and library scope'] broken=[] unresolved=[]
Add Product To Cart: links=['Add Product To Cart', 'Carts and library scope', 'Get Cart Item Count'] broken=[] unresolved=[]
Get Cart Item Count: links=['Add Product To Cart', 'Assertions', 'Get Cart Item Count'] broken=[] unresolved=[]
Get Cart Total: links=['Add Product To Cart', 'Assertions', 'Get Cart Total'] broken=[] unresolved=[]
Get Product Count: links=['Assertions', 'Get Product Count'] broken=[] unresolved=[]
Get Product Price: links=['Assertions', 'Get Product Price'] broken=[] unresolved=[]

Agent:

Libdoc lists all five keywords. Every Assertions link resolves, and so does every other link. Marking 6.1 done, then moving to the README.

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

Agent:

Working on task 8.1: README usage section.

The agent reads <repo>/atest/resources/demoshop.resource

The agent reads <repo>/atest/cart_keywords.robot

The agent writes <repo>/README.md

File created successfully at: <repo>/README.md (file state is current in yo

Agent:

Now I'll check the README against the libdoc signatures and run its example test against the shop.

The agent runs uv run python - <<'EOF' import json, re html = open("docs/DemoShopLibrary.html", encoding="utf-8").read() i = html.index("libdoc = ") + len("libdoc = ") data, _ = json.JSONDecoder().raw_decode(html[i:]) readme = open("README.md", encoding="utf-8").read() rows = dict(re.findall(r"^\| ([^]+) | (.+?) |", readme, re.M)) rows.pop("Keyword", None) assert set(rows) == {k["name"] for k in data["keywords"]}, rows.keys() for kw in data["keywords"]: libdoc_args = [a["name"] + (f"={a['defaultValue']}" if a["defaultValue"] is not None else "") for a in kw["args"]] readme_args = re.findall(r"([^]+)", rows[kw["name"]]) print(("OK " if libdoc_args == readme_args else "DIFF"), kw["name"], libdoc_args, readme_args) init = [a["name"] for a in data["inits"][0]["args"]] print("init args:", init) EOF mkdir -p /tmp/readme-check && awk '/```robotframework/{f=1;next} /```/{f=0} f' README.md > /tmp/readme-check/readme_example.robot && cat /tmp/readme-check/readme_example.robot && uv run robot --outputdir /tmp/readme-check /tmp/readme-check/readme_example.robot | tail -8

OK   Add Product To Cart ['product_id', 'quantity=1'] ['product_id', 'quantity=1']
OK   Get Cart Item Count ['assertion_operator=None', 'assertion_expected=0', 'message=None'] ['assertion_operator=None', 'assertion_expected=0', 'message=None']
OK   Get Cart Total ['assertion_operator=None', 'assertion_expected=0', 'message=None'] ['assertion_operator=None', 'assertion_expected=0', 'message=None']
OK   Get Product Count ['assertion_operator=None', 'assertion_expected=0', 'message=None'] ['assertion_operator=None', 'assertion_expected=0', 'message=None']
OK   Get Product Price ['product_id', 'assertion_operator=None', 'assertion_expected=0', 'message=None'] ['product_id', 'assertion_operator=None', 'assertion_expected=0', 'message=None']
init args: ['url', 'space']
*** Settings ***
Library    demoshop_library.DemoShopLibrary    url=http://localhost:9090    space=team-7

*** Test Cases ***
Buy Two Desks
    Get Product Price      5    ==    799
... (11 more lines)

Agent:

The README signatures match libdoc, and its example test passes against the shop. Marking 8.1 done and starting the final full run (8.2).

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent runs echo "=== pytest" && uv run pytest -q 2>&1 | tail -3 && echo "=== robot" && uv run robot --outputdir results atest 2>&1 | grep -E "^(Atest|[0-9]+ tests?,)|FAIL" ; echo "robot exit=${PIPESTATUS[0]}"

=== pytest
........................................................................ [ 94%]
....                                                                     [100%]
76 passed in 0.13s
=== robot
Atest                                                                         
Atest.Cart Keywords :: Cart keywords. Every test starts with an empty cart ...
Atest.Cart Keywords :: Cart keywords. Every test starts with an em... | PASS |
7 tests, 7 passed, 0 failed
Atest.Getter Assertions :: The assertion arguments shared by every Get keyw...
Atest.Getter Assertions :: The assertion arguments shared by every... | PASS |
8 tests, 8 passed, 0 failed
... (9 more lines)

Agent:

Unit tests (76) and acceptance tests (26) pass. Next I'll run the 7.6 keyword-owner check and inspect the built wheel.

*The agent runs `uv run python - <<'EOF' from collections import Counter from robot.api import ExecutionResult, ResultVisitor

ALLOWED = {"demoshop_library.DemoShopLibrary", "ShopA", "ShopB"}

class Owners(ResultVisitor): def init(self): self.owners = Counter() self.bad = [] def start_keyword(self, kw): self.owners[kw.owner] += 1 if kw.owner not in ALLOWED: self.bad.append((kw.parent.full_name if hasattr(kw.parent, "full_name") else kw.parent, kw.full_name))

result = ExecutionResult("results/output.xml") v = Owners() result.visit(v) print("stats:", result.statistics.total.passed, "passed,", result.statistics.total.failed, "failed") print("owners:", dict(v.owners)) print("violations:", v.bad or "none") EOF echo "=== build"; rm -rf dist && uv build 2>&1 | tail -3 && uv run python -c " import zipfile, glob whl = glob.glob('dist/.whl')[0] names = zipfile.ZipFile(whl).namelist() print(whl); [print(' ', n) for n in names] need = [f'demoshop_library/{m}.py' for m in ('init','library','client','products','cart')] + ['demoshop_library/py.typed'] print('missing:', [n for n in need if n not in names] or 'none') print('tests/atest present:', [n for n in names if n.startswith(('tests/','atest/'))] or 'none') "`

stats: 26 passed, 0 failed
owners: {'demoshop_library.DemoShopLibrary': 54, 'ShopA': 9, 'ShopB': 5}
violations: none
=== build
Building wheel from source distribution (uv build backend)...
Successfully built dist/demoshop_library-0.1.0.tar.gz
Successfully built dist/demoshop_library-0.1.0-py3-none-any.whl
dist/demoshop_library-0.1.0-py3-none-any.whl
   demoshop_library/
   demoshop_library/__init__.py
   demoshop_library/assertions.py
   demoshop_library/cart.py
... (10 more lines)

Agent:

Every check in 8.2 passes. Marking it done.

The agent edits <repo>/openspec/changes/add-demoshop-library/tasks.md

The file <repo>/openspec/changes/add-demoshop-library/tasks.md has been upd

The agent runs openspec instructions apply --change "add-demoshop-library" --json | uv run python -c "import json,sys; d=json.load(sys.stdin); print(d['state'], d['progress'])"; openspec validate add-demoshop-library 2>&1 | tail -2

all_done {'total': 22, 'complete': 22, 'remaining': 0}
Change 'add-demoshop-library' is valid

Agent:

Implementation Complete

Change: add-demoshop-library Schema: spec-driven Progress: All 22 tasks are done, and openspec validate reports the change as valid.

Completed This Session

6.1 Keyword documentation: No changes were needed; the docstrings already followed the reference shape. I checked the generated docs/DemoShopLibrary.html: it lists all five keywords, every Get keyword's assertion_operator row links to Assertions, and every other link works too.
8.1 README: I wrote a usage section in README.md with an import example, a table of the five keywords, and the commands for unit tests, acceptance tests, libdoc and build. The table's arguments match the libdoc output, and the README's example test passed against the shop.
8.2 Final run:
- uv run pytest: 76 passed.
- uv run robot --outputdir results atest: 26 passed, 0 failed.
- The 7.6 check found every executed keyword owned by demoshop_library.DemoShopLibrary, ShopA or ShopB.
- uv build: the wheel contains the required modules and py.typed, and no tests/ or atest/.

Two things to fix in the plan before archiving

The spec's libdoc command fails. The "Generate documentation" scenario in specs/library-import/spec.md uses robot.libdoc demoshop_library.DemoShopLibrary ... without the URL. That fails with expected 1 to 2 arguments, got 0, because url is required. Task 6.1 and AGENTS.md use the form with ::url=http://localhost:9090, which works, and that's the form I used. The project context returned by the OpenSpec CLI has the same broken command. I'd update the spec scenario to the ::url= form (/opsx:update) rather than make url optional, which would go against the import requirement.
An extra module isn't in the design. src/demoshop_library/assertions.py was added in an earlier session. It rejects text operators before any request is sent, which task 4.2 needs. The package layout in design.md (decision 1) doesn't list it. One line in the design would fix that.

Once those are updated, you can archive with /opsx:archive.

The participant runs grep -c '^- \[x\]' openspec/changes/*/tasks.md; grep '^- \[ \]' openspec/changes/*/tasks.md || echo 'no open task':

22
no open task

The participant runs uv run pytest uv run robot --outputdir results atest uv run python -m robot.libdoc "demoshop_library.DemoShopLibrary::url=http://localhost:9090" docs/DemoShopLibrary.html uv build:

------------------------------------------------------------------------------
Two Instances Have Separate Carts                                     | PASS |
------------------------------------------------------------------------------
Test Adds Products                                                    | PASS |
------------------------------------------------------------------------------
Cart Does Not Carry Over To The Next Test                             | PASS |
------------------------------------------------------------------------------
Test Setup Shares The Test's Cart                                     | PASS |
------------------------------------------------------------------------------
Suite Setup Cart Is Not Visible To Tests                              | PASS |
------------------------------------------------------------------------------
Atest.Library Import :: Importing the library: URL, workshop space... | PASS |
... (28 more lines)

The participant runs `cat > results-assert.robot <<'ROBOT' *** Settings *** Library demoshop_library.DemoShopLibrary url=http://localhost:9090

*** Test Cases *** The price passes Get Product Price 1 == 249.99

The wrong price fails with AssertionEngine's message Run Keyword And Expect Error * Get Product Price 1 == 1 ${message}= Run Keyword And Expect Error * Get Product Price 1 == 1 Log To Console ${message} ROBOT uv run robot --outputdir results-assert results-assert.robot; rm -f results-assert.robot`:

==============================================================================
Results-Assert                                                                
==============================================================================
The price passes                                                      | PASS |
------------------------------------------------------------------------------
The wrong price fails with AssertionEngine's message                  Product 1 price '249.99' (float) should be '1.0' (float)
| PASS |
------------------------------------------------------------------------------
Results-Assert                                                        | PASS |
2 tests, 2 passed, 0 failed
==============================================================================
Output:  <repo>/results-assert/output.xml
... (2 more lines)

Step 8 - Compare it with user keywords

Prompt:

Compare this library with ../ai-engineering-robotframework/resources/api.resource, which reaches the same API with user keywords and RequestsLibrary. Name one thing the library gives that the resource does not, and one thing it costs. Three sentences at most.

The agent runs ls ../ai-engineering-robotframework/resources/ 2>&1; ls ../ai-engineering-robotframework/resources/ 2>&1; ls src/demoshop_library tests atest

api.resource
catalogue.resource
checkout.resource
legacy.resource
shop.resource
api.resource
catalogue.resource
checkout.resource
legacy.resource
shop.resource
atest:
cart_keywords.robot
... (23 more lines)

The agent runs cat ../ai-engineering-robotframework/resources/api.resource; echo ----; cat src/demoshop_library/library.py src/demoshop_library/client.py src/demoshop_library/cart.py src/demoshop_library/products.py src/demoshop_library/assertions.py

*** Settings ***
Documentation       The shop's API, in the configured workshop space.

Library             Collections
Library             RequestsLibrary


*** Keywords ***
Open Shop API
    [Documentation]    Creates the API session. It sends ${SHOP_SPACE} when one is set.
    &{headers}=    Create Dictionary
    IF    "${SHOP_SPACE}"
... (405 more lines)

Agent:

Gives: The library's keywords check the result in the same step, and their failures are worded in shop terms. For example, Get Product Price 1 == 249.99 compares two floats and fails with Getting product 999999 failed: Product not found (HTTP 404), whereas Get Catalogue From API returns raw JSON, so the test has to make its own checks and gets RequestsLibrary's generic HTTP error. Costs: The library is a Python package with its own unit tests, dependency pins, build and release, so supporting a new endpoint means writing Python, while the resource is about 25 lines of Robot that any tester can edit in place.

Step 1 - Start the project next to the clone​

Step 2 - Save the references​

Keywords to imitate, and AssertionEngine's code, from the workshop's environment

Step 3 - Write AGENTS.md​

Step 4 - Set up OpenSpec​

Step 5 - Propose the first slice​

Step 6 - Review the proposal​

Step 7 - Apply it​

Implementation paused​

Done this session​

Problem: libdoc fails with the command the plan prescribes​

Options​

Still to do once you decide​

Step 7 - Apply it, continued​

Implementation Complete​

Completed This Session​

Two things to fix in the plan before archiving​

Step 8 - Compare it with user keywords​

Step 1 - Start the project next to the clone

Step 2 - Save the references

Step 3 - Write AGENTS.md

Step 4 - Set up OpenSpec

Step 5 - Propose the first slice

Step 6 - Review the proposal

Step 7 - Apply it

Implementation paused

Done this session

Problem: libdoc fails with the command the plan prescribes

Options

Still to do once you decide

Step 7 - Apply it, continued

Implementation Complete

Completed This Session

Two things to fix in the plan before archiving

Step 8 - Compare it with user keywords